Mentoring is widely viewed as an effective preventive intervention for a variety of youth problems and consequences, including conduct disorder and delinquency, academic failure and school drop out, substance use and abuse, and early sexual behavior and teen parenthood. In recent years, there has been strong support for mentoring programs from policymakers at the local, state, and federal levels. While there have been a variety of outcome studies on mentoring, and mentoring appears to have at least small, positive effects on youth in the short run, little is known about the long term influence of programs on the "high risk" children assumed to need mentoring the most, and almost nothing is known about how characteristics and qualities of the mentoring relationship are related to youth outcomes. We propose to conduct a longitudinal randomized controlled preventive intervention trial designed to assess the impacts of an established, privately funded, long term, intensive youth mentoring program called Friends of the Children (FOTC). FOTC partners with inner city public schools to identify and enroll kindergarten-age children thought to be at high risk for conduct disorder, academic failure, and teen parenthood due to serious levels and/or combinations of individual and family risk factors. FOTC mentors, called "Friends", are paid, full time employees who are rigorously trained and supervised. Friends work with not more than eight children at a time. Through FOTC, children are exposed to a variety of skill building and recreational experiences and assisted in accessing resources in a variety of areas, including academic and physical and mental health services. Participants for the study (N = 256;80% minority, 50% girls) will be recruited through existing programs located within impoverished, inner city areas of Boston;San Francisco;Seattle;and Portland, Oregon. Children will be randomly assigned to either FOTC or a referral control condition. Children, their primary caregivers, and Friends will be assessed prior to program initiation in kindergarten, and during each subsequent year through the end of third grade. On an annual basis, primary teachers will be assessed;direct observations of Friend-child social interaction will be conducted;and school and FOTC program records will be collected. At each assessment, information will be collected on child behavioral, emotional, and academic adjustment, as well as on the quality and characteristics of child-peer, child-caregiver relations, child-Friend, and child-other adult (i.e., natural mentor) relationships. Analyses will examine child outcomes overtime;the impact of Friend-child relationship quality on program satisfaction, engagement, and persistence as well as on child outcomes;and differential effectiveness based on ethnicity, child gender, and baseline risk status. The findings from this study will have important implications for mentoring practice and policy.